LFCT Validation Suite: Quantitative Predictions vs Observational Data (Data Runs E, F, G, H)

This archive contains six independent validation runs for Light Frame Cadence Theory (LFCT), refreshed against the post-Session-62 corpus state (2026-06-04). Each run computes quantitative predictions from LFCT's axiom set A1–A4 using zero free parameters and compares them against published observational data. The bundle is a six-folder zip with a top-level README documenting scope, framework versions, scoring convention, per-run summaries, and observational sources. All scripts are self-contained Python (no external data files); each run prints a full scorecard to stdout and exports machine-readable JSON + human-readable Markdown results. Scope (75 tests across six runs): Run D — What LFCT Explains That Standard Cosmology Cannot (explanatory-power validation: dark matter, dark energy, Hubble tension, MOND, CMB anomalies; 14 tests, score 0.93). Run E — LFCT's Numbers vs the Sky (quantitative number-vs-number validation of the master derivation chain; 13 tests, score 0.98). Run F — LFIS-27 Predictions (CMB architecture, peak scaffold, electroweak crossover; 8 tests, score 1.00). Run G — Slip, Polarisation, and Geometry (gravitational slip, ISW, EE/TT, B-mode, horizon geometry; 7 tests, score 0.96). Run H — Nine-Correction CMB TT Production Model (full TT spectrum 2 < ℓ < 2508 from the cadence-star limaçon; HRMS 0.94%, TrRMS 0.94%, score 1.88%, all 15 features within ±2%). Run I — Cross-Domain Precision Validation, Post-Session-62 (α, α_G, m_H/m_P, T_P/T_CMB, R_NS, magic-number nuclear architecture, a_0 cross-scale identity, D(M) closure ceiling, Hawking-T family; 18 tests, score 96.7%). Headline cross-domain results include 1/α = 137.0359990780 matching CODATA at 0.044 ppb, α_G = 1.7517×10⁻⁴⁵ at ≈1.8σ_G consistency, and m_H/m_P at ~0.5% — all from structural composition with zero fitted parameters. Framework versions validated against: LFCT Foundation Axioms v2.2.0 (concept DOI 10.5281/zenodo.20306153), Core Conceptual Spine v4.0.3, Core Mathematical Framework v4.0.1, Core Physical Derivations v3.0.4, LFIS-24 v2.14.4 (master spine), LFIS-29 v1.6.2 (CMB TT capstone + a_0 cross-scale architecture), LFIS-31 v1.13.2 (nuclear binding), LFIS-34 v1.1.1, LFIS-36 v2.0.0, Notation Guide v2.16.9, and the full LFIS publish-bundle. All structural constants in the runs (c = π, κ_TD = 1, κ_TS = π², κ_TR = 5/2, ε = 1/π², C₀ = 1/π, g_s = 5/7, D(M) = 2M+3) are derived from axioms A1–A4, not fitted. Changes from v1 (2026-04-05 deposit): (1) Added Run I covering the ~18 precision-claim derivations that entered the corpus between v1 release and Session-62 publish-bundle close. (2) Refreshed Runs D–H against the current framework versions listed above. (3) G4 EE/TT normalisation fix: the corpus-locked formula is now EE/TT = 25/[4π⁴(2π²+1)] = (κ_TR/κ_TS)² · r_TR via the cadence-overhead factor r_TR = 1/(2π²+1), per LFIS-24 v2.14.4 §D26.399 + Core PD v3.0.4. This replaces the earlier 25/(8π⁶) = (κ_TR/κ_TS)² · ε/2 form (a prior corpus state with r_TR = ε/2). G4 discrepancy improved from 6.6% to 1.5% (PASS); BB/TT propagated to 25/[6π⁸(2π²+1)]. The Run G v1 script is preserved unchanged in the original v1 deposit; pre-fix copy archived locally. This suite is the machine-process complement to the Publication Audit Checklist §10 Numerical-Sanity Sweep (full-corpus pass completed 2026-06-04); concordance between the two is part of LFCT's zero-free-parameter discipline.